Clinical trials are in the middle of a technology- enabled transformation.
Smartphones and wearables are lessening, and in some cases eliminating, the need for subjects to visit study sites. At the same time, the data explosion and emergence of automated analytical tools have raised expectations that the industry can crack the patient enrolment conundrum and other persistent problems. If sponsors can clear the remaining bottlenecks, the era of end-to- end digital drug development will begin. That era has been a long time coming. Researchers used computers to collect and analyse patient data back in the 1960s and, after a slow start, sponsors have moved en masse to electronic clinical trial technologies over the past 20 years. These advances failed to transform trials, though. Digital replacements for paper-based processes made parts of drug development more efficient, but the resulting technology-enabled clinical trials had the same structure – and, by extension, the same shortcomings – as their flawed predecessors.
False dawn
The ongoing challenges created by these flaws are illustrated by data. Around 2% of eligible people in the US take part in clinical trials, according to the Center for Information & Study on Clinical Research Participation. This makes it hard for sites to enrol. Calculations of the proportion of sites that never enrol a patient range from around 10% to 30%. Non- enrolling sites add to costs and trial durations. The problems continue once patients are enrolled. With the typical trial requiring participants to visit a site 11 times in six months, a significant minority of patients drop out before the end of the study.
A recent UK study put the average drop-out rate at 11% but it can be far higher. GlaxoSmithKline lost 40% of the subjects in a diabetes trial, making it impossible to draw hard conclusions from the data. Clinical trial burdens affect study investigators, too. Around 40% of FDA-registered investigators stop running clinical trials every year, according to the Tufts Center for the Study of Drug Development. The high drop- out rate stops the system from accruing a pool of experienced, fast-enrolling investigators. Technology was supposed to make investigators’ lives easier and their trials more efficient. Yet, while technology helped in some ways, it also created new challenges and failed to address core problems. That experience should temper expectations about the impact of technology, but there are reasons to think the current changes are fundamentally different from the series of advances that brought us to this point.
End-to-end digital trials
Today, the industry is in the middle of an end- to-end digital reimagining of trials that may mark the biggest change to clinical research in decades. Clinical trials performed before and after the adoption of electronic data capture (EDC) were the same in many key respects. The same cannot be said the adoption of technologies that will limit, and in some cases end, the need for physical study centres.
Digital technologies are starting to reshape each step in the clinical trial process, from study design to closeout. In the first steps in the process, sponsors are factoring real-world evidence into trial designs and using models to forecast outcomes. Beyond that, there are opportunities to use data analytics to select the sites most likely to enrol patients, digitise screening and informed consent and optimise the supply chain with electronic technologies.
At the end of the trial process, sponsors can automate data analysis and the generation of tables and graphs, cutting the time it takes to understand results and enabling companies to cope with the fast growth in data sets. Companies are experimenting with such automation. “We developed some machine learning, some simple statistical data science algorithms in the cloud,” Dirk Schapeler, director of digital heath at Bayer, said at an event in 2017. “There was no special equipment needed. The whole computing power to do that is actually not something that we needed to buy. It’s just something that you can actually tap into in the cloud.”
Still bigger breakthroughs may come from technologies applied to the conduct of clinical trials. It is at this stage in the process that smartphone apps, wearables and connected devices are redefining data capture and patient monitoring. Before connected devices, investigators gathered data from patients during periodic site visits and made conclusions based on these snapshots. The health of subjects between visits was unknown. Connected devices are plugging that gap by collecting patient data around the clock, giving sponsors unprecedented insights into the effects of their therapies. This approach has multiple benefits. While in-clinic tests show how a patient was on a particular hour, wearables give a more comprehensive picture, reducing the risk that outlying data points will skew the results. This is important given evidence that vital sign data collected in clinics and in the real world differs.
The phenomenon is common enough to have a name: white-coat hypertension. A version of the phenomenon affects trials in muscle disorders such as Duchenne muscular dystrophy (DMD). These trials assess efficacy by having subjects perform the six-minute walk test. However, the results are affected by how participants feel at the time of the test. One study found kids with DMD walked 11% further if they were offered $50 (£39) to beat their previous distance. Using ankle-worn wearables to calculate stride velocity at home could eliminate the confounding effect of motivation. In these and other ways, sensor-laden wearables are enabling the remote assessment of safety and efficacy. They are far from the only connected devices used by sponsors, though. Another set of technologies monitors and encourages adherence to treatment regimens.
To provide such insights, Oracle integrated Proteus Digital Health’s adherence-tracking smart pill into its EDC system, while other companies use connected tablet bottles, app notifications and SMS text messages to encourage adherence. Connected inhalers provide comparable oversight of adherence for inhaled medications. Going virtual Individually, these technologies and many others in development stand to improve distinct parts of clinical trials. Some of these improvements could be large enough to notably impact the efficiency of clinical research but the biggest gains will come from harnessing multiple technologies in tandem. Most dramatically, the availability of digital informed consent tools, supply chain sensors, wearables and other devices is enabling sponsors to eliminate or significantly reduce the need for study sites and in-person visits.
These virtual and hybrid designs use technologies and processes to perform the tasks traditionally done at centres, enabling people to enrol in studies regardless of their location and without committing to an arduous schedule of site visits. Attracted by the potential to speed enrolment and slash costs, Genentech, Novartis, Pfizer and other leading drugmakers have adopted virtual and hybrid designs. However, while home visits by nurses can do some tasks that technology cannot, many protocols require procedures that must still be performed at sites, such as MRIs. Consequently, it is more common for sponsors to use technology to reduce, but not eliminate, study centres and the frequency of site visits. These obstacles to fully virtual trials look set to be long-lasting but other current barriers to the use of technology are likely to be temporary.
Many of the temporary barriers stem from the immaturity of technologies and the data sets supporting their use. For example, in September the European Medicines Agency provisionally ruled that more robust data is needed before the aforementioned DMD wearable is used to gather primary endpoint data. The agency thinks the idea is sound, though. As sponsors, technology companies and regulators work through these barriers, the number of trials that can be fundamentally reimagined through the use of technology will increase. Sceptics and optimists may disagree about how long this will take but the direction of travel is set. End-to-end digital drug development is coming. The question is how soon.
